Streaming data has become more popular and useful in recent years due in part to improvements in data compression techniques, which has made the transmission of data via a network increasingly more feasible, and due in part to improvements in memory technology (e.g., storage and processing). Commercial video-content providers and video-sharing sites have helped to popularize the streaming of data.
While streaming of video content is one relatively popular use for streaming data, other types of data can be streamed as well. For example, it can be desirable to stream and/or store data representative of media content, such as video or audio content. One type of data that can be representative of video content is video (or data) patterns or structures (e.g., video fingerprints) that can be based on the video content. For example, unique features of digital video content can be extracted from the digital video content and can be stored as a video structure, such as a digital video fingerprint, that can be representative of and/or can facilitate identification of that digital video content. A digital video fingerprint can be highly compressed, as compared to the original digital video content, which can allow for lower storage and/or streaming costs than that associated with uncompressed or less compressed data.
While there have been improvements in technology relating to data streaming, there are still a number of challenges in being able to desirably stream data content, for example, via a network (e.g., the Internet). For example, there can be times when it can be desirable to take a repository containing data offline to perform general maintenance or repairs on the repository. Also, there can be instances where a network outage may occur, which can make the repository unavailable to a client requesting streaming of data. It can take time for a repository, or portion thereof, to recover after being made available again (e.g., after being put back online, after a network outage is rectified, etc.). The unavailability of and recovery time for a repository may cause an undesirable disruption to, or may otherwise negatively impact, streaming of the data due in part to, for example, low latency requirements that may be associated with data streaming.